As far as I can tell, that is only one piece of the puzzle. I think the bulk of the work is on the hypervisor side (KVM) to eliminate/stub out sources of randomness in HW, patch out RDRAND, etc